app.py

from fastapi import FastAPI
from pydantic import BaseModel
from transformers import LlamaForCausalLM, LlamaTokenizer
import torch

app = FastAPI()

model = LlamaForCausalLM.from_pretrained("memorease/memorease-quizgen")
tokenizer = LlamaTokenizer.from_pretrained("memorease/memorease-quizgen")

class Memory(BaseModel):
    description: str

@app.post("/generate")
def generate(memory: Memory):
    prompt = f"Soru üret: {memory.description}"
    inputs = tokenizer(prompt, return_tensors="pt", padding=True, truncation=True, max_length=128)
    outputs = model.generate(**inputs, max_new_tokens=64)
    question = tokenizer.decode(outputs[0], skip_special_tokens=True)
    return {"question": question}